|
|
| GeoCommunity Mailing List |
| |
| Mailing List Archives |
| Subject: | RE: GISList: OGC and Standards, - a response |
| Date: |
01/07/2003 11:20:53 AM |
| From: |
Dimitri Rotow |
|
|
>
> After reading through all the posts, I am dismayed that this
> discussion has
> deteriorated to a debate about data formats instead of exploring
> the issues
> that started this thread.
If I remember correctly, the issues that started this thread revolved around
an assertion that the practical effect of OGC is to codify obsolete
approaches to GIS to protect the obsolete product technologies of legacy
vendors, while also playing a supporting role in codifying an architectural
approach favored by institutional members who do not really want to release
their public GIS holdings into the hands of the public.
The discussion of GML format I believe is not really about the GML format
but exploring a specific example case of how OGC specs end up being deformed
and standing in the way of modern GIS. It's a typical OGC "carrier pigeon"
technology that by it's very nature prevents a rapid and effective modern
techonology.
>
> In my opinion as a former sponsor and current vendor, OGC's current thrust
> in in distributed geoprocessing via the Interoperability Program is not
> about "the average GIS user." In fact the "average GIS user" is
> a miniscule
> part of the folks that use spatial data on a daily basis. OGC is about
> making geospatial processing available to everyone, not just technical
> specialists with desktop systems.
>
Well, I appreciate your candor. Now, if OGC was honest as you are about
their objectives they would make it clear on their website that those of use
who are technical specialists with desktop systems are *not* the
constituency served by OGC and that OGC will happily craft specs that reduce
our quality of life in favor of chasing the rainbow of "making geospatial
processing available to everyone." That would save us technical specialists
with desktop systems from wasting any time thinking that OGC specs are
something we would want to get involved with.
The idea of "making geospatial processing available to everyone" involves a
number of least-common-denominator constraints. In the real world, a
utopian approach of crafting an architecture that thinks it can provide
geospatial processing to everyone will result in compromises that prevent
the high-performance desktop from being all it can be.
Technically, there is really a choice: either you design for a high
performance, interactive, modern desktop or you design for an arms-length,
low-performance web interface. But, you cannot optimize both. There are
people who say you can use the same interfaces within a high performance
desktop that can also serve the low-performance solution. It's true that
you can, but if your desktop solution is held back compromises made for a
more low performance architecture you will never have as fast, powerful,
modern and elegant high performance solution as can be created by someone
who is designing for the high performance desktop without constraints.
Much of what one sees come out of OGC is just the old idea of "doing GIS
over a network" come back from the dead. Why this is foolish if one wants a
high performance desktop experience is laid out in the essay at (watch the
line wraps):
http://exchange.manifold.net/manifold/manuals/5_userman/mfd50GIS_and_Network
ing.htm
>
> The first message is the concept of the "geospatial dial-tone." The idea
> behind the the geospatial dial-tone is that like a 2600 Hz signal either a
> person or a machine knows what to do request information from a
> service and
> what to do with the information that the service provides. I was recently
That's fine as a general aspiration, but if it is implemented in a way that
works against high performance computing it is not such a good idea.
> driving in Europe and called a business associate in the US on my US cell
> phone. I was amazed at the technology that connected me to a foreign cell
> carrier, sent my call across the Atlantic, and negotiated the different
> networks to reach another cell phone roaming in the US. What makes this
> possible are service level contracts based on international standards that
> tie together loosely coupled information systems. OGC is
A poor analogy.
The international phone network started life as an analog network that was
nothing more complex than a big set of switches with standards largely
driven by the power of a single monopoly, AT&T. The international standards
organizations for the most part simply rubber stamped what AT&T had already
made a standard.
If that's the model you think is right for GIS, then quit OGC and go to work
for Microsoft and let them set the standards through .NET and similar
Microsoft technologies just as AT&T did for telephones.
> striving for this
> level of transparency for geospatial processing and it is succeeding. In
> numerous testbed
|
|
Sponsored by:
For information
regarding
advertising rates
Click Here!
|
